Roadrunner's AI-Native Ambition to Fix Enterprise Quoting

The labyrinthine world of enterprise sales, long burdened by archaic software and cumbersome processes, is ripe for disruption by artificial intelligence. This is the core thesis driving Joubin Mirzadegan, Kleiner Perkins’ latest entrepreneurial force, as he embarks on his new venture, Roadrunner. In a recent interview with Swyx on the Latent Space podcast, Mirzadegan peeled back the layers of his journey, from building go-to-market strategies at startups to incubating breakout companies like Glean, culminating in the launch of Roadrunner – an AI-native reimagining of Configure, Price, Quote (CPQ) workflows.

Mirzadegan, a seasoned sales leader turned VC and now founder, spoke with Swyx about the profound inefficiencies embedded in current enterprise quoting systems, a problem he intimately experienced throughout his career. "Probably the number one thing that used to break my back was that the underlying software with like Salesforce CPQ and others, just to like create a quote, get it approved, is horrific," Mirzadegan stated, painting a vivid picture of sales teams grappling with 30-second loading screens in the critical final days of a quarter. This widespread frustration, he argues, presents a colossal opportunity for a ground-up rebuild.

The solution, according to Mirzadegan, lies in harnessing the power of large language models (LLMs) to abstract away the immense complexity inherent in enterprise pricing. His "lightbulb moment" came from observing how LLMs could reason with both structured and unstructured data in other domains. "I actually think you can abstract away a bunch of the complexity with these LLMs... and it's, you know, unstructured and structured text that you can reason with, do stuff with," he explained. This capability, already transforming fields like legal research with tools such as Harvey, can fundamentally alter how deal structures are recommended, policies enforced, and approvals routed, eliminating the "Slack-channel chaos of deal desk." Roadrunner's ambition is to build an AI-native architecture that not only streamlines quoting but also intelligently recommends optimal deal configurations, effectively acting as a digital deal desk.

Mirzadegan's insights are deeply rooted in his experience at Kleiner Perkins, where he helped technical founders navigate the often-treacherous waters of go-to-market (GTM) strategy. He highlights key lessons from successful incubations like Glean, an enterprise AI search leader now valued at $7 billion. A critical takeaway from his work with Glean's founder, Arvind Jain, was the realization that "go-to-market is not native to him." This underscores a pervasive challenge for technically brilliant founders: translating product edge into repeatable revenue. The success of companies like Windsurf, which Mirzadegan also supported, hinged on a dual commitment to "Google-class product and Salesforce-class distribution"—a rare but potent combination.

Hiring the right sales leadership early on is paramount, yet fraught with "anti-patterns." Technical founders often fall into the trap of recruiting based on "fancy logos" on a LinkedIn profile. However, Mirzadegan warns against this superficial approach. "The problem with that is that you have to actually interview them. Like, you can't just see what they did on their LinkedIn profile and know if they're good or not." He stresses the importance of digging deeper, understanding a candidate's intrinsic motivations, their "chip on their shoulder," and their willingness to operate as an "artist" in the early, unstructured stages of a startup, rather than simply executing a pre-defined playbook. Stage-fit, he emphasizes, is everything.

Beyond product and sales, Mirzadegan also reflected on the strategic use of media. His own podcast, "Grit," started as a deliberate "hiring wedge" and relationship engine to connect with top Chief Revenue Officers. His philosophy for content creation—to "commit to 100 episodes" before judging its success—speaks to the perseverance required to find one's voice and build an audience. He also noted a significant difference in interviewing founders versus other executives. "Founders and CEOs have an authority to speak in a different way than somebody on the executive team... they can just speak." This candor, unburdened by concerns about managerial approval, unlocks more earnest and honest conversations, providing invaluable insights into their entrepreneurial journeys.

Roadrunner’s vision is not merely to iterate on existing CPQ tools but to fundamentally reconstruct the data model from the ground up, co-developing with "hairiest design partners." This approach aims to leverage LLMs to recommend complex deal structures, enforce policies, and automate approval routing in a way that legacy systems, with their brittle "rules + SKUs + approvals" graph, simply cannot. The two-year window, Mirzadegan posits, is critical: incumbents are rebuilding slowly, while AI-native startups like Roadrunner have the agility to out-sprint them.

Mirzadegan’s journey and the inception of Roadrunner exemplify a deep understanding of market pain, technological potential, and the human elements of building a successful enterprise. His experiences highlight the necessity of not just building exceptional technology, but also mastering the art of distribution and assembling a team driven by genuine motivation and stage-appropriate expertise.